stm32f4xx_hal_i2s_ex seems incomplete - doesn't work

mike2399 · ‎2016-07-08

Posted on July 08, 2016 at 18:04

I'm developing an audio application using STM32F427, using I2S port 3 in ''full-duplex'' mode to interface to ADC and DAC.

Trying to use the non-blocking-interrupt method, as I don't want to use DMA and there are several related functions in the same RTOS thread that I don't want to block.

On trying to trace the interrupt handling code, as far as I can see, the function I2SEx_TransmitReceive_IT doesn't contain any callback calls in the I2S_MODE_MASTER_TX or I2S_MODE_SLAVE_TX section. There are callbacks in the RX section, but they leave the peripheral in HAL_I2S_STATE_BUSY - which will cause any call to HAL_I2SEx_TransmitReceive_IT (in order to re-start transmission/reception) to fail. It also leaves the macro __HAL_LOCK(hi2s) in the locked state.

I have searched through this forum history and can't see any solution. I did see a related question, but the poster got round it by using DMA.

I'm using code originally created by CUBE FW.F4_V1.11.0, which uses HAL version 1.4.4. I downloaded the latest CUBE (version 1.12), with HAL version 1.5.0 - but the included code is identical.

Any chance of some fully complete and working HAL files? Or have I missed something?

Amel NASRI · ‎2016-07-15

Posted on July 15, 2016 at 17:08

Hi fleetwood.mike,

Thanks for highlighting this issue. It is reported internally to be fixed in one of coming releases of the STM32CubeF4 package.

-Mayla-

To give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.

ingwmeier · ‎2016-07-17

Posted on July 18, 2016 at 07:57

Same issue with an STM32F051 app.

We used HAL_I2C_Master_Transmit / HAL_I2C_Master_Receive in the end to read 20bytes from a AS5600 magnetic encoder chip. This works fine in the beginning. After several readings an extra 0x00 byte is in the reception buffer at start which is not visible using a scope. We have no time to investigate this at this moment, so we use a patch to ignore this byte since the result of the chip is never 0x00 at start..

Amel NASRI · ‎2016-07-18

Posted on July 18, 2016 at 11:44

Hi Werner Meier,

This is an I2S related issue/topic (not I2C).

Could you please start a new thread in which you describe exactly the problem you are facing? If it is already done, please provide the link.

-Mayla-

To give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.

ingwmeier · ‎2016-07-18

Posted on July 18, 2016 at 15:22

Ups sorry..

for I2C in interrupt mode, it worked one time,

then stayed also got stuck in HAL_I2C_STATE_BUSY mode.. I will open a new topic

mike2399 · ‎2016-08-01

Posted on August 01, 2016 at 14:10

Many thanks!

Since then, I have re-written the I2S in/out in low-level code, reading/writing registers direct. But I'm sure the updated HAL code will be useful to someone.

Mike.